Inorg. Chem. 1984, 23, 2223-2227 Titration data, the apparent absence of H2evolution, and the demonstration of a 3: 1 redox stoichiometry (N-chloromorpholine serves as a two-electron oxidant of I-) have been interpreted as evidence for essentially quantitative oxidation of all hydridic hydrogen by hypochlorite. Stoichiometric and kinetic studies of the reaction of NaOCL with MeiNH.BH3 (Table IV) suggest such a pattern of reactivity to be generally characteristic of secondary amine-boranes. The question is raised, then, as to whether tertiary amine-boranes react by a distinctly different pathway from secondary amine-boranes or whether chloroboranes are formed via a single mechanism in both systems, in which case observed stoichiometric and kinetic differences reflect differences in (subsequent) reactivity of tertiary and secondary amine-chloroboranes. Implication of HOCl as the reagent involved in rate-limiting attack seems to be a common feature in the reactions of both tertiary and secondary amine-boranes. If different mechanisms are involved, then, necessarily, the mode of attack by HOCl on the borane adducts must be different. A speculative model of a (four-center) activated complex leading to the formation of chloroboraneamine is depicted in I and involves U
I R,, R,, R, =alkyl
an orientation of OH and C1 with respect to a B-H bond opposite to what has been suggested for HOCl attack on Mbn.' Defense of such a model requires one to accept the notion that when Rl, R2, or R3= H, the favored pathway involves attack by oxygen of the HOCl molecule (rather than chlorine) at boron. This might be considered a reflection of differences in the capacity of N-bonded substituents (alkyl vs. hydrogen) to transmit electron density inductively to the boron center. Alternatively, the presence of an N-H group might permit a hydrogen-bonded interaction with HOCl that would stabilize an activated complex favoring 0-atom attack on boron. Subsequent attack on amine-monochloroboranes and amine-
2223
dichloroboranes would be expected to occur at progressively slower rates due to the electron-withdrawing influence of boron-bonded chlorine. The alternate view, that chloroboranes are formed in either case, requires acceptance of the premise that, unlike the case with tertiary amine-boranes, successive chlorination of secondary amineboranes proceeds progressively faster as more chlorine is introduced into the coordination sphere of boron. It also requires stepwise chlorination from R2NH.BH3 through R2NH.BC13 to occur quantiatively without hydrolytic loss of hydridic hydrogen, followed by rapid hydrolysis of the aminetrichloroborane to produce free secondary amine, which will be subject to rapid chlorination. Though not disproven, this scheme seems unlikely in view of the reported instability of secondary amine-haloboranes in aqueous media.I4 Of the two alternatives, we tend to favor that involving separate transition-state configurations. Admittedly, there are disadvantagous features to a scheme that suggests an apparently subtle change from N-H to N-alkyl to cause an alteration in reaction pathway as opposed to a change in the rate of a comparable transformation proceeding by a common pathway. On the other hand, the B-H bond is not inherently highly polarized and it may not be unreasonable to imagine N-alkyl substitution to alter the directional polarization of a transition-state configuration, particularly in view of the well-known changes effected by B-substitution on reaction mechanism in the hydrolysis of amineboranesg,'0,'3,15-'8 as well as the effect of N-substitution in establishing the relative importance of acid-independent and acid-dependent pathways for hydrolysis of alkylamine-, arylamine-, and selected heterocyclic amine-borane~.~J~J~ Acknowledgment. This research was supported by the Robert A. Welch Foundation (Grant P-162). We also acknowledge assistance of the TCU Research Fund. (14) Niedenzu, K.; Dawson, J. W. "Boron-Nitrogen Compounds"; Academic Press: New York, 1965; pp 27-42. (15) Weidig, C.; Uppal, S. S.;Kelly, H. C. Inorg. Chem. 1974, 13, 1763. (16) Weidig, C.; Lakovits, J.; Kelly, H. C. Inorg. Chem. 1976, 15, 1783. (17) Skillern, K.; Kelly, H. C. Inorg. Chem. 1977, 16, 3000. (18) Kelly, H.C.; Marriott, V. B. Inorg. Chem. 1979, 18, 2875.
Contribution from the Department of Chemistry, Boston College, Chestnut Hill, Massachusetts 02 167
Equilibria and Kinetics of Complexation of Bidentate Ligands with the Macrocyclic Complex Ni([14]aneNJ2+ E. J. BILL0 Received July 26, 1983
The kinetics of reaction of several bidentate ligands X (ethylenediamine, glycinate, oxalate) with planar Ni( [ 14]aneN4)2+ to produce cis octahedral Ni([ 14]aneN4)(X)has been examined. Both forward and reverse rate constants have been determined. The forward reaction is first order in nickel complex, independent of [XI over the range 2-40 mM, and first order in [OH-]; the rate law is kf = 13[OH-] s-'. The reverse reaction is first order in nickel complex, first order in [OH-], and inverse first order in [XI; the rate law is 2.6 X 10-2[OH-]/[X] s-I. A mechanism is proposed for the reaction. In the forward direction, the folding of the macrocycle (base-catalyzed trans cis isomerization) is rate determining, followed by rapid coordination of the bidentate ligand. In the reverse direction, rapid preequilibrium dissociation of the bidentate ligand is followed by rate-determining cis trans isomerization. This mechanism is in sharp contrast to a recent study of the reaction of bidentate ligand with the related complex Ni(trien)z+. Stability constants for the reaction of X with Ni( [ 14]aneN#+ are reported.
-
Introduction The kinetic stability of nickel(I1) macrocycle complexes is strongly influenced by the macrocycle ring size, with the saturated tetraaza macrocycle [ 14]aneN, (1,4,8,1 l-tetraazacyclotetradecane) providing a complex of remarkable kinetic
+
inertness toward dissociation.' Tetradentate macrocyclic ligands exhibit an overriding tendency to coordinate to metal ions in a planar geometry. This is especially true in the case (1) Busch, D.H.Acc. Chem. Res. 1978, 11, 392.
0020-1669/84/ 1323-2223%01.50/0 0 1984 American Chemical Society
2224 Inorganic Chemistry, Vol. 23, No. 15, 1984
Billo
of nickel(I1) complexes, where the in-plane ligand field exertkd by the macrocycle stabilizes low-spin, square-planar nickel(I1). The formation and dissociation reactions of macrocyclic complexes, however, must involve intermediates in which the ligand is coordinated in a folded, cis geometry. Such intemediates have been postulated,2 but little kinetic evidence has been provided. Recently, a paper from our laboratory described a facile preparative path to the folded complex cis-Ni( [ 141aneN4)(H20)22+and reported a kinetic study of the foldedto-planar isomeri~ation.~In the present study, the kinetics of reaction of bidentate ligands (X = en, Gly-, ox2-) with the planar Ni( [ 14]aneN4)2+complex (eq 1) is examined. Ni([14]aneN4)*+
+X
G=
Ni([l41aneNd(X)
kf, k,, K (1)
The study was undertaken in order to obtain kinetic and equilibrium information about the reverse isomerization process, from the planar to the folded geometry. Before the experiments were performed, it was assumed that the reaction would be first order in the attacking bidentate ligand as has been found for a large number of other systems4 Somewhat surprisingly, it was found that the rate is independent of the concentration of the bidentate ligand. The rate step of the reaction is determined by the macrocycle. This is in sharp contrast to a recent study by Steinhaus and Lee,5 on the reaction of bidentate ligands with the related complex Ni(trien)2+ (trien = H2N(CH2)2NH(CH2)2NH(CH2)2NH2). Experimental Section Reagents. Ni( [ 14]aneN4)(C104)2was synthesized by the method of Barefield6 and recrystallized several times from water. Glycine and sodium oxalate were reagent grade and were used as received. Ethylenediamine was redistilled; a stock solution was prepared by weight and standardized by pH titration. Kinetic Measurements. The rate of formation of the mixed-ligand complex was measured by observing the disappearance of planar Ni([14]aneN4)2+at 450 nm, by using a Cary 14 spectrophotometer and thermostated IO-cm cells (initial concentrations: Ni( [ 14]aneN4)*+, 1.0-2.2 mM; X, 2.0-40.5 mM). Rate constants were obtained from plots of In ( A - A , ) vs. time. All rates were measured at 25 OC; ionic strength was adjusted to 0.10 by addition of sodium perchlorate. Reaction pH was adjusted by using borate buffer or by using the bidentate ligands (en or Gly) as buffers. No difference was detectable in reactions with and without borate buffer. Hydronium and hydroxide ion concentrations were obtained from -log [H’] = pH - 0.11 and log Kw = -13.78 for 25 O C , I = 0.10.
Results The stability constants of the mixed-ligand complexes are relatively low, owing to the planar tendency of the macrocycle. As a result, with ligand concentrations of 2-40 mM, particularly at the lower pH values, the reaction did not proceed to completion but rather to a position of equilibrium. From each kinetic run, therefore, the stability constant of the mixed-ligand complex could be calculated in addition to the rate constant, by using the equilibrium absorbance A,. Stability Constants. The equilibrium constant for reaction 1 was calculated by using eq 2-4 (L = [14]aneN4; X = en, Gly-, ox2-). Charges are omitted for clarity. [NIL] = A,/bt
(3)
(4) . (5) (6)
(2)
Busch, D. H.; Farmery, K.; Goedken, V.; Katovic, V.; Melnyk, A. C.; Sperati, C. R.; Tokel, N. ACS Monogr. 1971, No. 100, 44. Ekstrom, A,; Leong, A. J.; Lindoy, L. F.; Rodger, A,; Harrison, B.; Tregloan, P. A. Inorg. Chem. 1983, 22, 1404 and references therein. Billo, E. J. Inorg. Chem. 1981, 20, 4019. Margerum, D. W.; Weatherburn, D. C . ;Cayley, G.; Pagenkopf, G. K. ACS Monogr. 1978, No. 174. Steinhaus, R. K.; Lee, B. I. Inorg. Chem. 1982, 21, 1829. Barefield, E. K.; Wagner, F.; Hodges, K. D. Inorg. Chem. 1976, 15, 1370.
Table I. Stability Constants for Binding of Bidentate Ligands to Nickel(1I) Centers log K”
reaction
+ + ’++
Ni(H,O),” en Ni(H,O),” GlyNi(H,O), oxzNi(en),(H,0)22+ e n Ni(trien)(H,O),” en Ni(trien)(H,0),2+ GlyNi([ 14]aneN,)(H,0)2Z’ en Ni([14]aneN,)(H,0),Zt Glyox2Ni([ 14]aneN,)(H,0),z’ cis-Ni([ 14]aneN,)(H,0),2+ t en cis-Ni([ 14]aneN,)(H,0),zc + Glycis-Ni( [ 14]aneN,)(H,0),2’ ox2-
+ + +
+ + +
+
7.6
5.9
5.3 4.3 5.3b 4 . 0 , ‘ ~4.34 ~ t 0.02e 2.56 t 0.05f3g 2.75 t O.lOfSg 2.6 t 0.2f,g 4.2d,f 4.4d3f 4.2dpf
’From ref 8, unless otherwise noted.
Reference 4. Reference 5 . From kinetic measurements. e Israeli, Y. J.: NiL Volpe, R. Bull. SOC.Chim.Fr. 1971, 3119. This work. is a mixture of planar ( 7 0 % ) and octahedral (30%).
The molar absorptivity of planar Ni( [ 14]aneNJ2+ depends on the concentration of perchlorate ion.’ The following values of t, determined in this study, were used in the calculations ([C104-]/M, t/M-’ cm-I): 0.002, 38; 0.01, 40; 0.03, 42; 0.05, 45; 0.10, 50. Resolution of the spectra of the mixed-ligand complexes into Gaussian bands showed that there was no significant contribution of the mixed-ligand species to the absorbance at 450 nm. [NiL], - A , / b € K= (3 1 ( A , / b € ) ( [ x ] ~- [NiL], A,/bt)a
a = (1
+ KIH[H+]+ KIHK2H[H’]2)-’
(4) In eq 3 and 4 [NiLIT and [XI, are the analytical concentrations and K,Hand K2Hare the protonation constants of the ligands. Equation 4 calculates the fraction of ligand in the complexing form. For en, log KIH= 10.0 and log KzH= 7.0; for Gly-, log KIH = 9.6.8 All experiments with oxalate were performed at pH 9.62 f 0.03, buffered with 0.05 M borate. The protonation constants of oxalate are small, and no corrections for protonated ligand are necessary at the pH values used in this study. The stability constants obtained are listed in Table I with, for comparison, the corresponding stability constants of the ligands with aquonickel(I1). The value for the en complex is in excellent agreement with the earlier value (log K = 2.6 f 0.1) determined under conditions of no added NaC104.4 The similarity between the results for the three ligands is surprising. The stability constants are essentially identical, within experimental error, while the values for reaction with aquonickel(I1) illustrate the much stronger binding constant of a nitrogen donor compared to a carboxylate. There is no doubt, however, that the mixed-ligand complexes are as formulated in eq 1. The spectra of the product solution in each case is in agreement with the expected nickel(I1) chromophore, and in the case of the en complex, C,H,N, analysis4 of the readily isolated violet precipitate is in agreement with the formula Ni( C10H24N4) (C2H8N2)(C104)2. Kinetics of Formation of NiLX. Under the experimental conditions used (small K, excess [XI), reaction 1 is reversible and first order or pseudo first order in each direction. Plots of In ( A - A , ) vs. t under conditions of rever~ibility~ yield koM values given by eq 5 . Since k f / k , = K, the forward and reverse (5) kobsd = kf -k k r (7) Fabbrizzi, L. Inorg. Chim. Acta 1977, 24, L21. (8) Sillen, L. G.; Martell, A. E. “Stability Constants of Metal-Ion Complexes”; The Chemical Society: London, 1964; Supplement No. 1, 1971. (9) Moore, J. W.; Pearson, R. G. ‘Kinetics and Mechanisms”, 3rd ed., Wiley: New York, 1981; p 304.
Inorganic Chemistry, Vol. 23, No. 15, 1984 2225
Bidentate Ligand Complexes with Ni( [ 14]a11eN,)~+
I
HZO
I
RAPID
I
A--8
'
+AB
A
Figure 2. Two possible mechanisms for the reaction of a bidentate ligand (here shown as AB) with planar Ni([14]ar1eN,)~+.
8
10
9
-log
II
[ H*]
Figure 1. Variation of log kfas a function of -log [H'] ([XI = 2, 4, 10, 20, 40, mM;25 OC;I = 0.10 M (NaCIO,)): circles, glycinate; squares, ethylenediamine; diamonds, oxalate. Open symbols indicate reactions run in 0.050 M borate buffer; solid symbols indicate those run without borate buffer. The line is calculated by using eq 6 .
components of koMcan be determined. The extent of reaction R = [NiLX]/[NiL], can be readily calculated from the expression R = (Ao- A , ) / A O ,from which the individual rate constants can be obtained: kf = Rkobsd,k, = (1 - R)kobsd. Plots of In ( A - A,) vs t were linear over several half-lives, even when [XI, [NiLIT, indicating a reaction zero order in X. After correction for reversibility, kf was found to be independent of [XI and base catalyzed (Figure 1). Within experimental error, runs using 2,4, 10, 20, and 40 mM ligand concentrations fall on the same line when log kf is plotted vs. -log [H+], As well, results from all three ligands are identical. The line has a least-squares slope of 1.04 f 0.02 and correlation coefficient of 0.997. The rate law is given by eq 6.
-
kf = 13[OH-]
(6)
In view of the observed ligand independence and base catalysis of the formation reaction, a mechanism is proposed in which the rate-determining step is isomerization of the macrocycle from planar to folded, followed by rapid coordination of the bidentate ligand (eq 7). An alternative mechanism planar
slow
folded
X
NiLX
(7)
would be the one suggested by S t e i n h a d for the analogous reaction of Ni(trien)2+ + X (X = 1,lO-phenanthroline, glycine). In this mechanism (eq 8), a monodentate intermediate, NiL
+X
rapid
NiL-X*
slow
NiLX
(8) Ni(trien)-X* in rapid equilibrium with reactants is formed, followed by rate-determining ring closure. This mechanism predicts first-order dependence on X at low concentrations, shifting to zero-order dependence on X at high [XI when the intermediate is fully formed. For the Ni( [14]aneN4)2+system, this mechanism is discarded on the basis of the following arguments: (i) No deviation from zero order is detected, even at the lowest ligand concentrations. (ii) Zero-order dependence on [XI requires complete formation of the monodentate intermediate, yet estimates of the equilibrium constant for monodentate coordination of a nitrogen donor of either ethylenediamine or glycine1° to a Ni(I1) tetraaza complex
suggest that, even at the highest concentration of X used, less than 50% of the intermediate complex would be formed. (iii) The proposed intermediate would be octahedral or at least more than four-coordinate, yet when absorbance data from kinetic runs was extrapolated to t = 0, there was no detectable decrease in Ao, from that calculated for the square-planar complex NiLZ+,by using the molar absorptivities given above. Formation of a few percent of an octahedral intermediate would have been detected. The two possible paths for reaction are shown in Figure 2. An additional factor to be considered in the Ni([ 14]a11eN,)~+ system is the extremely rapid low-spin-high-spin equilibrium K,ll (at 25 OC, 0.10 M NaClO,, K1 = 0.4 (70% low-spin NiL2+, 30% high-spin tran~-NiL(H~O),~'). Consider first the mechanism proposed by Steinhaus5 for the reaction of Ni(trien)2+ with glycine. This mechanism, rapid formation of the intermediate followed by rate-determining ring closure, when applied to the Ni( [ 14]ar1eN,)~+case, would be the lower path shown in Figure 2. The slow ring closure, perhaps 100 times slower than "normal" ring closure rates, was attributed by Steinhaus for the Ni(trier~)~+ case to steric hindrance by the trien ligand. Spectral evidence12 indicates that the trien complex is already in the cis geometry, ci~-Ni(trien)(H,O)~~+, and so the actual path for the trien reaction is simply the rapid coordination of the ligand to the folded complex, shown as the single vertical step on the right in Figure 2. If this mechanism were to be applied in the present case, the NiL complex would be planar (trans-NiL(H20)22+)and the slow ring closure would be the result of steric and kinetic effects of the concomitant folding of the macrocycle. Second, consider the alternate path, initial isomerization of the planar to the folded form followed by rapid coordination of X. Cassatt and WilkinsI3 measured the rate of reaction of a number of ligands with aquonickel(I1); for glycinate, the rate constant is approximately 2 X lo4 M-' s?. This will be used here as an estimate of the rate of reaction of Gly- with cisNi([14]aneN4)(Hz0)2+. In an earlier study4 in our laboratory, the (reverse) isomerization rate of cis-Ni( [ 141aneN4)(Hz0)pto the trans form was measured; it is also base catalyzed and has the rate law (9). It was estimated that k2
k-2 = 560[OH-]
(9)
(10) Estimated by using the equation of Jackobs and Margerum (Jackobs, N. E.; Margerum, D. W. Inorg. Chem. 1967,6,2038) from which log K = 1.9 or by using log K5for Ni2+-NH3 (log K5= 0.9) and including a statistical factor of 5, from which log K = 1.6. Coupled with K, = 0.4, the overall K would be ca. 20. This estimate does not include the steric interactions expected from the macrocycle. (11) Wilkins, R. G.; Yelin, R.; Margerum, D. W.; Weatherburn, D. C. J . Am. Chem. Soc. 1969, 91,4326, Pell, R. J.; Dodgen, H. W.; Hunt, J. P. Inorg. Chem. 1983, 22, 529. (12) Cook, D. F.; McKenzie, E. D. Inorg. Chim. Acta 1978, 31, 59. (13) Cassatt, J. C.; Wilkins, R. G . J . Am. Chem. SOC.1968, 90, 6045.
2226 Inorganic Chemistry, Vol. 23, No. IS, 1984
Billo contribution to kobsd.) The rate expression is given by eq 12,
I -5.
-
25
I
L
s
0,
0
-6
,
-7
9
IO
II
- l o g [H']
Figure 3. Variation of k , [ X ] as a function of -log [H']. For the meaning of symbols, see Figure 1. The line is calculated by using eq 12.
is at least 100 times slower. Thus, even at the lowest Glyconcentration used, the subsequent coordination of X is much faster than base-catalyzed isomerization. These estimates confirm that macrocycle folding can indeed be the rate-determining step. The rate and equilibrium constants involved in the proposed mechanism are defined in (lo). The ligand-independent, K
Ni( [14]aneN4)2+
k
5 k-i
trans-Ni( [ 14]aneN4)(H20)22+ ki(X1
cis-Ni( [ 141aneN,) (H 2 0 )22+ '
k-3
Ni( [ 141aneN,) (X) (10)
base-catalyzed formation rate kf is give by eq 11. Since K, kf = Klk,[OH-]
(1 1)
= 0.4, k2 is calculated to be 30 M-l s-l. This cis-trans equilibrium constant can thus be calculated from the kinetic data: K2 = k2/k-,! = 301560 = 0.055. From KI (determined spectrophotometrically) and K2, the percentages of planar, trans, and cis in aqueous solution at equilibrium (25 ' C , I = 0.1) are calculated to be 69% 29% and 2% respectively. From the measured equilibrium constant for the formation of NiLX and the rate and equilibrium constants calculated from kinetic data, the equilibrium constant K3 for the reaction of bidentate ligand with cis-Ni( [ 14]aneN4)(H2022+can be calculated. Thus, K(measd) = K1K2K3,and hence K3 = k-2K/Klk2. From data for ethylenediamine (log K = 2.56 f 0.05, K = 360 M-'), a value of K3 = 1.6 X lo4 (log K3 = 4.2) is obtained. Similarly, from the stability constant values of glycine and oxalate, log K3 values of 4.4 and 4.2, respectively, are obtained. These values can be compared with some literature values in Table I. Kinetics of the Reverse Reaction. Since the rate law for the forward reaction does not contain [XI, the reverse rate law is expected to be inverse first order in [XI. A plot of log k,[X] vs. -log [H+] was linear (Figure 3) with slope 0.92 f 0.04 (correlation coefficient 0.98) while a plot of log k, vs. -log [H'] was scattered (correlation coefficient 0.64). The unit slope shows that the reverse reaction is also base catalyzed. (The deviation of the least-squares line from 1.OO, the larger error in the slope, and the smaller correlation coefficient for the reverse reaction, compared to the forward reaction, reflect the fact that, for many of the rates measured, k, was a minor
the data of Figure 3 yielding a value of 2.6 X s-I for k-2k-3/k3. As was observed for the forward reaction, the reverse reaction is independent of the identity of the ligand, at least within experimental error. Although this seems paradoxical at first glance (that the reverse rate is inverse first order in ligand concentration but independent of the identity of the ligand), it is simply another manifestation of the steric effect of the coordinated macrocycle on the binding of the bidentate ligand. This steric hindrance, particularly between NH2 groups on the bidentate ligands en and Gly and ring methylenes of the macrocycle, gave rise to similar stability constants for three different ligands. The reverse rate constant k, is a composite of k-2 (cis-trans isomerization of the macrocyclic complex) and K3 (the stability constant for binding of X, which has essentially the same value for all three of the ligands). As well, the first-order dependence on [OH-] tends to conceal small differences in the reverse rate, if indeed there are differences. From eq 12 and the previously measured rate constant k-2 for the folded to planar is~merization,~ the stability constant K3 can be calculated. Since the reverse rate constant was the same for the three ligands studied, only a single value of K3 can be calculated. A value of K3 = 2.2 X lo4 M-l (log K3 = 4.3) is obtained, in excellent agreement with the values found in the forward rate study.
Conclusions The results presented above are consistent with a mechanism in which rearrangement of the macrocycle geometry is the rate-determining step. Rearrangement from planar to folded requires inversion of two of the coordinated nitrogens.I4 Base catalysis of the rearrangement is assumed to occur via the conjugate base mechanism proposed for cobalt(II1) complexes.ls Hydroxide catalysis of the inversion of nitrogen atoms in the related, copper(I1) complex of 5,5,7,12,12,14hexamethyl- 1,4,8,11-tetraazacyclotetradecane has been reported.I6 The rearrangement occurs by means of coordinated hydroxide. A similar mechanism cannot be ruled out at this time. It is interesting to compare the rate laws and mechanisms for the substitution of bidentate ligands at three different nickel(I1) centers: Ni(H20)2*, Ni(trien)2+,Ni( [ 14]aneN4)2+. In the case of hexaaquonickel(II), the rate law is k[Ni2+][XI and the mechanism is the familiar Eigen-Wilkins water loss mechanism. The rate of substitution is limited by the rate of loss of coordinated water from Ni(H20)62+.Substitution of glycine on Ni(trien)2+ should be faster than on Ni(H20)62+, since Margerum has shown that coordinated nitrogens increase the lability of remaining coordinated water molecules. Substitution is found to be slower, however, and the results of Steinhaus and Lee demonstrate that in the case of a highly substituted nickel(I1) center, chelate-ring closure can be rate determining, leading to a rate law in which the order in the entering ligand decreases from first order toward zero order. In contrast to this second study, however, the present work provides an example in which rearrangement of the tetraaza ligand, prior to coordination of the bidentate ligand, is rate determining. The high-spin-low-spin equilibrium and (presumably) steric crowding of the remaining coordination sites by the macrocycle combine to lower the stability constant for (14) Whimp, P. 0.;Bailey, M. F.;Curtis, N. F. J . Chem. Soc. A 1970, 1956. (15) Buckingham, D. A,; Marzilli, L. G.; Sargeson, A. M. J . Am. Chem. Soc. 1967, 89, 825. (16) Liang, B.-F.; Margerum, D. W.; Chung, C.-S. Inorg. Chem. 1979,18, 2001.
2227
Inorg. Chem. 1984, 23, 2221-2231 monodentate binding of the ligand X,so that the alternate path (proposed by Steinhaus for the Ni(trien)2+ Gly reaction) is not favorable. It is likely that other nickel(I1) tetraaza ligand complexes that have either planar or trans-diaquo geometries and therefore must undergo rearrangement of the tetraaza
+
ligand in order to react with polydentate ligands will react by a mechanism similar to the one proposed in this study. Registry No. en, 107-15-3;Ni([ 14]aneN,)2+,46365-93-9; glycine, 56-40-6; oxalic acid, 144-62-7.
Contribution from the Departments of Chemistry, Smith College, Northampton, Massachusetts 01 063, and University of California, San Diego, La Jolla, California 92093
Equilibrium and Kinetic Studies of Substitution Reactions of Fe(TIM)XY2+ in Aqueous Solution ALISON BUTLERla and R. G. LINCK*lb
Received April 25, 1983 Equilibrium and rate constants for substitution reactions of some bisligated complexes of Fe(T1M) (TIM is 2,3,9,10tetramethyl- 1,4,8,1l-tetraazacyclotetradeca-1,3,8,lO-tetraene) in aqueous acetonitrile solutions are reported at 23 "C and an ionic strength of 0.5 M. The equilibriumconstant for the replacement of CH3CN by H 2 0 (K,) in Fe(TIM)(CH,CN)22+ is 0.023 0.002 whereas the same replacement in Fe(TIM)(CO)CH3CN2+( 4 )is 1.1 f 0.1. The mechanism and the rates of establishment of these equilibria are dramatically affected by the *-accepting nature of the trans ligand, namely CH3CN or CO. Substitution trans to CH3CN is very rapid, with equilibration occurring on the time scale of milliseconds. On the other hand, the rate of approach to equilibrium in the replacement of CH3CN for HzO in Fe(TIM)(CO)CH3CN2+ is complicated by two paths, one in which the ligand trans to the CO is directly replaced and the other in which the CO leaves, equilibrationof the coordination sphere of the Fe(I1) center occurs, and CO then recoordinates. This rate of approach to equilibrium occurs with a rate constant dependent upon the [CH,CN], being about 3 X lo-, s-I at a [CH,CN] of 1.0 M. In delineation of the complexities of the rate of approach to the equilibrium described by K3, the rate constant for CO loss from Fe(TIM)(CO)H202+was determined to be 3.6 X s-l, and rate constants for all other steps involved were determined.
*
Introduction Since the synthesis of 2,3,9,10-tetramethyl-1,4,8,1 l-tetraazacyclotetradeca-1,3,8,1O-tetraene(hereafter abbreviated TIM) by Rose and co-workers in 1969,2 a number of interesting chemical and physical properties of Fe(TIM)XY"+ molecules have been rep~rted.~" These properties led us to consider the possibility that this iron complex could act as a sensitizer for a number of photolytic reactions in aqueous solution. This possibility is given some credence by the report of Rose and co-workers, who have established that the bis(methanol) adduct, Fe(TIM)(CH30H)?+, is oxidized by dioxygen in air and that the resultant Fe(II1) complex is then susceptible to photolysis that leads to oxidation of methanol to formaldehyde accompanied by regeneration of the original Fe(I1) ~ o m p l e x . Moreover, ~ our attention was drawn even more strongly to this molecule by the report of Incorvia and Zink,5 who found that photolyses of Fe(TIM)(CO)CH3CN2+ are different depending on the solvent in which the photolysis takes place, a result that they interpreted as a highly unusual solvent mediation of a primary photophysical event. We decided, therefore, that further study of this molecule was appropriate to learn more about both its thermal and photolytic properties. A variety of questions are raised by these considerations, excluding how the processes take place. Among them are the following: What, if any, are the substitutions that take place in Fe(TIM)XY2+in aqueous solutions? What is the time scale (a) Part of Ph.D. dissertation resea.rch, UCSD. (b) To whom correspondence should be addressed at Smith College. Baldwin, D.A.; Rose, N. J., paper presented at the 157th National Meeting of the American Chemical Society, Minneapolis, MN, April 1969; see Abstracts, No. INOR 020. Baldwin, D.A.; Pfeiffer, R. M.; Reichgott, D.W.; Rose, N. J. J. Am. Chem. SOC.1973, 95, 5152. Reichgott, D.A.; Rose, N. J. J. Am. Chem. SOC.1977, 99, 1813. Incorvia, M. J.; Zink, J. I. Inorg. Chem. 1977, 16, 3161. Stynes, D.V.;James, B. R. J. Chem. SOC.,Chem. Commun. 1973,325.
0020-166918411323-2227$01 SO10
in which substitution reactions of the axial ligands take place? What is the affinity of the Fe(TIM)"+ center for axial ligands with back-bonding characteristics? What can be said of the inherent oxidation-reduction reactivity of Fe(TIM)XY2+? In this and following reports on this molecule we will address these questions. Our initial experiments on Fe(TIM)XY2+quickly established that substitution reactions occurred on two separate time scales, one on the order of less than a second and the other on the order of minutes to hours. In this paper we report equilibria occurring when Fe(TIM)(CH3CN)?+ and Fe(TIM)(CO)CH3CN2+are in aqueous acidic acetonitrile solution and the kinetics of the slower processes of substitutions. In future papers, we will discuss the rapid substitution reactions of Fe(TIM)(CH3CN)22+,Fe(TIM)(CH3CN)H202+,and Fe(TIM)(H20)2+ and results of oxidation-reduction kinetics of Fe(TIM)XY*+.*
Experimental Section Materials. [Fe(TIM)(CH,CN),](PF,), was synthesized according to the procedure of Reichgott and Rose: modified as indicated.Argon was used as an inert gas, and the reaction was stirred for 48 h at 35 OC under a stream of this gas rather than being sealed for 24 h under nitrogen. The crude yield contained a substantial amount of a tanbrown impurity, which was removed by recrystallization; about 0.5 g of the crude material was added to approximately 7 mL of CH,CN, and the mixture was filtered quickly. Cold ethanol (about 50 mL) was added, and the mixture was filtered as soon as possible. The burgundy-colored solid thus obtained was washed with cold ethanol and cold ether quickly, so as to avoid any oxidation. [Fe(TIM)(CO)CH,CN](PF,), was prepared by the published proced~re.~ Solutions of Fe(TIM)XYZ+were prepared from the solid material by weight or analyzed for iron content by utilization of the 550-nm peak of Fe(TIM)(CH3CN)22+,where the extinction coefficient in ~~~
~~~~
(7) Butler, A,; Linck, R.G. Inorg. Chem., in press. ( 8 ) Flores, M.; Linck, R. G., to be submitted for publication.
0 1984 American Chemical Society
